Influence of Electric Bicycle Usage on Biker Effort

On-road Monitoring Application in Lisbon, Portugal

Magno Mendes, Gonçalo Duarte and Patrícia Baptista

IDMEC, Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais, 1, 1049-001 Lisboa, Portugal

Keywords: Physiological Signals, Heart Rate, Electric and Conventional Bicycles Monitoring.

Abstract: Bicycle use in urban environments is an alternative mobility option, which enables people to travel longer,

faster and with less effort than walking, with low environmental impacts. The use of electric bicycles (EB)

has risen as another possibility to promote a more efficient transportation use. However, the quantification

of the real impacts for the biker of shifting from conventional (CB) to EB is not yet quantified. This

research work aims at estimating the impacts on physiological signals, namely, on heart rate, from using EB

instead of CB, using a suitable methodology for on-road bio-signals data analysis. The on-road monitoring

of 6 bikers, 2 routes and 3 bicycles in Lisbon presented a 57% average reduction in HR variation from using

EB, since under high power demanding situations, the electric motor attenuates human effort. It was also

possible to estimate the energy expenditure associated to the human effort that results from using the

bicycles. For the CB the total energy spent reaches 70 Wh/km, while the EB presents 51 Wh/km of

human energy (28% lower than the CB) and 9 Wh/km of electricity consumption, resulting in a total of

60 Wh/km. Consequently, the total energy per km is 14% lower in the EB compared to the CB.

1 INTRODUCTION

The transportation sector faces increasingly

demanding energy consumption and emissions

standards representing 33% of the final energy

consumption, with the road transportation sector

being responsible in 2011 for 82% of that energy

consumption (EUROSTAT, 2013).

One alternative to reduce the impact of the

transportation sector, particularly in urban

environments, is to decrease the demand for energy

intensive modes of transportation and by promoting

alternatives that can provide a cheaper, less noisy

and more sustainable alternative than a day-to-day

car commute. Generally, three alternative

transportation models can be identified: public

transportation systems (bus, trains, subway systems

and others), vehicle sharing schemes (such as cars or

bicycles), and alternative transportation modes such

as walking, private bicycles or others (Wang, 2011).

From these different alternatives, the use of bicycles

is one of the more advantageous as it allows the

users to move at significant speeds for short

distances (typical in urban environments), resulting

in no emissions and having health benefits (Lindsay

et al., 2011).

Using bicycles enables people to travel longer,

faster and with less effort than walking, while

having a low impact on environment, thus making it

an efficient transportation mode for urban mobility.

As a result, the importance of cycling has been

increasing worldwide (Freemark, 2010; Urban

Audit). In many developing countries, namely in

Asia, two-wheelers are a first affordable step

towards individual mobility. In European and

American cities, the deployment of city bike ways

infrastructure has also been increasing, with bike

sharing systems deployed having in average 200 km

of bike lanes (Baptista, 2013).

Concerning the use of bicycles on urban

environments, a growing number of cities have been

trying to integrate them in the daily mobility of their

citizens, which for some countries has resulted in a

significant share of trips being done using a bicycle,

such as the Netherlands (26%), Denmark (18%) and

Germany (10%) (Buehler and Pucher, 2012). In the

city of Amsterdam, 38% of all trips in 2008 were

made using a bicycle, with 50% of Amsterdam’s

residents riding a bike daily and 85% riding it at

least once a week (Gardner, 2010).

While the use of conventional bicycles in an

urban context has been promoted with significant

success in several cities, namely Paris and London

with 25000 and 8000 deployed bicycles respectively

265

Mendes M., Duarte G. and Baptista P..

Inﬂuence of Electric Bicycle Usage on Biker Effort - On-road Monitoring Application in Lisbon, Portugal.

DOI: 10.5220/0004701202650274

In Proceedings of the International Conference on Physiological Computing Systems (PhyCS-2014), pages 265-274

ISBN: 978-989-758-006-2

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

(Barclays Cycle Hire, 2012); (Lathia et al., 2012);

(Vélib, 2012), they still have several drawbacks that

hinder their widespread use. Some of the main

problems identified by people when using

conventional bicycles for urban transportation

include the difficulty to travel very long distances

and over hills, the possibility of arriving at a

destination, such as work, sweaty or tired (Dill and

Rose, 2012), being exposed to extreme cold or hot

climates, among others. Even some cities with

difficult topographies, such as Lisbon, have also

begun promoting the use of bicycles through the

creation of bike lanes and studying the possibility of

having bike sharing schemes (Galp Energia, 2012);

(Martinez et al., 2012).

Several of these issues can also be overcome

through the use of electric bicycles (Dill and Rose,

2012). Generally, it is expected that the use of

electric bicycles can help reduce the effort required

for performing trips as well as reduce travel time,

though at a higher cost due to the electric system and

the energy used.

One of the main applications of electric bicycles

is in bike sharing systems, with several systems

being deployed worldwide. The Callabike system in

Aachen, Germany, has recently deployed a fully

electric bike sharing system with 15 electric bicycles

(Callabike, 2012). The city of Kitakyushu in Japan

also presents a full electric system with 116 bikes

(The Bike-sharing Blog, 2011). Cities such as St.

Etienne and Poitiers in France present mixed

conventional and electric bike sharing systems with

a 15% and 26% ratio between electric and

conventional bicycles respectively (Cap'Vélo, 2012);

(VéliVert, 2012).

Both conventional and electric bicycles are

starting to be seen as a real option under urban

environments, however, the real impacts on human

efforts have not yet been accounted under real

operation. Also, despite the high expectations for

electric bicycles, very few studies have tried to

understand the real world benefits of such bicycles

in an urban environment.

Regarding environmental impacts, for instance,

in China the estimation of environmental impacts

comparing electric bicycles with other means of

transport (bus) (Cherry et al., 2009) remarks that

electric bikes, in a life cycle perspective which

includes the well-to-tank stage, have higher

emissions of SO

(due to burning coal for electricity

production) compared to a bus, however the

emissions of other pollutants are lower in electric

bike. As result, pollutant emissions are strongly

related to the energy mix. The emissions associated

with the production process of batteries, recycling

and "dump" are also a concern.

Considering the adoption of electric bicycles, the

benefits of using electrical technologies are not

unanimous (Cherry and Cervero, 2007). The

potential environmental impacts, interference with

traffic and safety issues, as well as the potential

conflict between users of electric bikes and

conventional is a concern, since the speed

differences during cycling can pose a problem (Dill

and Rose, 2012).

Therefore, in terms of conventional and electric

bicycle usage comparison (Baptista et al., 2013b), a

16% increase in average speed was verified in

electric bicycle over that achieved with the

conventional bicycle. Different usage strategies of

the bicycle were also identified: the first strategy of

using the electric bike is to use a high level of

electric assist on positive slopes (uphill conditions),

lowering the electric assistance levels for neutral and

negative slopes; in the second strategy, the rider uses

more electric assist on the positive slopes, assistance

decreases in negative slopes, and reaching the lowest

values in the plain areas; and the third strategy is to

always use a high level of service regardless of the

slope.

The biker driving dynamics represented by the

speed and acceleration, combined with road

topography, reflects in a power demand that must be

overcome either by the biker (in a CB) or by the

biker and/or the electric motor (in an EB). The

quantification of human effort during cycling can be

addressed (Parkin, 2011), using a formula that

includes variables such as speed, acceleration,

mechanical efficiency of the bicycle, among others.

The author states that the slope of the road

influences the energy spent by the cyclist, as well as

the number of stops. Just stopping at an intersection

can lead to an increase of 10% in energy

consumption.

An important issue is the quantification of the

amount of effort or energy that the rider expends to

complete a specific route. More importantly,

whether an electric bike will actually decrease the

effort or energy expended by the rider when

compared with a conventional bicycle is also an

issue. There is little work developed in this field and

it does not reflect real world use of conventional and

electric bicycles. Therefore, a method to estimate

human energy expenditure (EE) must be addressed.

This method must include physiological data that

can be related with energy expenditure and an

analysis that could use on-road, real-world operation

of electric and conventional bicycles.

PhyCS2014-InternationalConferenceonPhysiologicalComputingSystems

266

A strategy for quantifying the human effort is the

comparison between ventilation and heart rate as an

indicator of oxygen consumption during exercise

with different intensities (Gastinger et al., 2010). By

monitor individuals performing different tasks (such

as walking, walking carrying a certain load and

intermittent work), the authors concluded that the

most appropriate methodology is the heart rate to

determine the oxygen consumption. Another

possibility is a calorimeter indirect way versus heart

rate monitoring to evaluate energy consumption (Yu

et al., 2012). The authors state that for determining

energy consumption, in tasks of day-to-day, the two

methods have very similar values, 8.6 kcal/day.

However, the authors argue that the method of using

the heart rate, to determine the energy consumed in

the daily tasks, still needs improvements.

Accurate estimate of energy consumption

through the heart rate without individual calibration

laboratory can be performed (Pulkkinen et al.,

2005). The authors argue that the methodology

IEST

where individual calibration of heart rate is

not necessary, provides an accurate and practical

way to estimate the power consumption.

The comparison of two techniques to estimate

the energy consumed, obtained by monitoring the

heart rate, and obtained by a portable

electromagnetic coil (Gastinger et al., 2012) allows

concluding that the determination of the energy

consumed using electromagnetic coil portable

system is more accurate than using only the heart

rate. The authors also report that would be

interesting to use together, heart rate and ventilation

on the determination of the energy consumed.

Several other techniques to determine the energy

consumed are available or being developed, with

particular reference to Doubly Labelled Water

(DLW) (Ainslie et al., 2003). In this study, the

authors argue that the methods used to determine the

power consumption depends on factors such as the

number of individuals monitored and the monitoring

period. The authors suggest that studies with few

participants and short analysis periods, should use

the method of indirect calorimeter to obtain best

results. However, for longer periods, around 3 to 4

days, it is preferable to use the method of DLW.

Although there are several techniques available,

the prediction of the energy consumed during

submaximal exercises could be done using heart rate

readings (Keytel et al., 2005). Through tests

conducted at 115 individuals in ergonomic bikes and

treadmills race, the authors established an equation

to determine the energy consumed by an individual

during exercise. This equation includes the

following variables: heart rate, age, sex and weight.

The authors claim that it is possible to determine

with good precision, the energy consumed using

only heart rate, age, sex and weight, and without the

need for individual calibration.

As can be seen, most of the techniques to

estimate human energy expenditure were performed

under controlled conditions – unlike the study

presented, but present solutions and correlations that

include signals such as heart rate that can be

collected while the bicycles are used.

According to this framework, the objectives of

this research work were to develop a methodology

based on physiological data collection under regular

bicycle operation. The goal was to evaluate the

application of conventional and electric bicycles for

urban mobility focusing on typical hilly routes of

Lisbon, quantifying their correspondent effect on

human energy expenditure.

2 METHODOLOGY

2.1 On-road Monitoring

The evaluation of electric and conventional bicycles

was done through the monitoring of trips performed

by 6 male different bikers (within the same age

range and physical characteristics), with each biker

travelling the same urban tour with both bicycles.

The bikers used the electric bicycle first and the

conventional bicycle after, with a minimum resting

period of 1 hour in-between.

The bicycles used by all bikers were the same, in

order to enable a fairer evaluation, although two

models of electric bicycles were evaluated. The

specifications of the three bicycles used are the

following:

 Conventional bicycle (CB) (Orbita Aluminio):

weight of 15 kg, 21 gears;





Electric bicycle (EB1) (QWIC Trend2): power

assist electric bicycle with six levels of

assistance, 25.7 kg, 7 mechanical gears and a

detachable Li-ion battery with a 360 Wh

capacity, provided by Prio.Energy (Prio Energy,

2012); and





Electric bicycle (EB2) (Ekoway L1): power

assist/power on demand electric bicycle with

23 kg, 6 mechanical gears and a detachable Li-

ion battery with a 360 Wh capacity, provided by

EcoCritério (Eco-critério, 2013).



Each trip was monitored during the ride using a

monitoring laboratory designed to assess energy and

InfluenceofElectricBicycleUsageonBikerEffort-On-roadMonitoringApplicationinLisbon,Portugal

267

environmental impacts associated to non-motorized

modes, MoveLab. This laboratory was assembled by

the DTEA - Transport, Energy and Environment

research group of IDMEC – IST and corresponds to

a backpack weighting 12 kg that the user (pedestrian

or biker) carries, as shown in Figure 1.

b) c)

Figure 1: MoveLab components and experimental

apparatus used for the real time monitoring (a), electric

bicycle (b) and conventional bicycle (c).

MoveLab is equipped with a GPS to record the

dynamic profile of the trip (including location,

altitude and speed), voltage and current probes to

assess the levels of electric assistance, and biometric

sensors (recording heart rate and breathing

intensity). All these equipment was carried by the

rider in the backpack. When asked to carry the

MoveLab backpack, the bikers saw no

inconvenience since in their daily routines they

already carry backpacks weighting around 5 to 8 kg.

All the MoveLab equipment is connected to a

laptop running a purposely developed software in

LabView to synchronize and record the data at 1 Hz,

throughout the trip. The technical description of the

equipment used is presented in Table 1.

Adding to the MoveLab, it was used bioPLUX

Research hardware and software acquiring

physiological data simultaneously. In order to

synchronize the two sources of data (LabView and

bioPLUX Research), a force sensor was adapted to

the numeric pad. This way, when the biker pressed

Enter, this signal was recorded either in LabView

and Plux software. Since the bio-signals require a

high frequency data logging, a minimum of 200

samples per second were collected while the biker

was riding. Due to the different temporal resolution

physiologic data was post processed into a second

by second time basis.

Table 1: Technical description of the equipment in

MoveLab.

Monitoring

equipment

Data acquired

Temporal

resolution of

data (Hz)

GPS (Garmin GPS

map 76CSx)

Speed (km/h), altitude

(m), location

Voltage and current

probes (Fluke i1010)

Voltage, current 1

bioPLUX Research

Heart rate, Breathing

rate

200

2.2 Data Collection and Processing

The GPS allows collecting speed, location and also

altitude information via an integrated barometric

altimeter. The altimeter was adequately installed

inside the backpack, avoiding pressure fluctuations

due to movement that could affect the readings. An

external antenna was used to avoid GPS signal

losses. Voltage probes were installed directly in the

bicycle battery terminals, while current

measurements were done on the circuit that connects

the battery to the electric motor. The signals

provided by the probes were collected by a National

Instruments DAQ board installed also on the

backpack. For battery voltage signal a voltage

divider circuit was placed before the DAQ board to

account for the 0-10 V limit of the acquisition

device. Both GPS and battery data were collected in

a PC using a program developed in LabView by the

authors to integrate the different communication

protocols (serial port and NMEA protocol for GPS

and analog data via USB port for the voltage and

current collected in the DAQ board) that allows to

synchronize the data, capturing all the equipment

readings in a 1 Hz basis. A solid-state disk PC was

used to avoid data loss while in motion.

The bioPLUX Research tool was used to collect

heart rate and breathing rate. This information was

collected at 200 Hz using PLUX software. Post

processing of data included the conversion of the

data to 1 Hz basis. It should be noticed that

breathing rate measured is very sensible to vibration

under regular bicycle operation, therefore it was

decided not to use this data.

PhyCS2014-InternationalConferenceonPhysiologicalComputingSystems

268

GPS readings of speed were used to post process

distance travelled, acceleration and road grade.

Altitude and distance were used to determine the trip

road grade using an algorithm that, for each point of

the trip, finds the points 50 m before and after and

uses this information to establish a second order

polynomial fit based on three points of distance and

altitude. The derivative of the polynomial fit in the

studied point allows determining road grade, which

is presented in rad.

Battery data was used to determine, at each point

of the trip, the power provided by battery to the

electric motor, according to the biker demands, and

integrate this data along the trip to find the

cumulative energy spent on the predefined tour.

With the data collected it is possible to

understand and quantify how riders changed their

use profile (in terms of speed and acceleration),

changing from a conventional bicycle to an electric

one. Also, the physiological impacts on the adoption

of an electric bicycle versus the conventional were

addressed using the physiological data, which also is

intended to provide an estimate of the human energy

expenditure.

2.3 Monitored Tours

To compare the use of conventional and electric

bicycles, round-trip tours of approximately 8.5 km

and 5.7 km were performed by each biker, in

Lisbon, with both bicycles. One of the tours

consisted on going from Instituto Superior Técnico

(IST) main campus to downtown Lisbon and back,

passing through the top of the Parque Eduardo VII

and Avenida da Liberdade on both ways. With this

tour, the bikers crossed different parts of the city of

Lisbon including traffic intensive avenues, side

roads with very little traffic and a street with a bike

lane. The other tour was carried in the EXPO 98

area, simulating a journey in a leisure place with low

traffic conditions. In terms of topography, the tours

had significant slopes, as summarized in Table 2.

Table 2: Selected routes.

Route

Distance

(km)

Average positive

slope (rad)

Average negative

slope (rad)

R1 8.54 0.037 -0.029

R2 5.66 0.020 -0.017

2.4 Methodology for Data Analysis

The analysis used in this work is based on Vehicle

Specific Power (VSP) to estimate the power demand

by vehicles, which combines speed (v), acceleration

(a) and road grade (). This methodology allows

comparing different technologies under similar

power requirements. It is traditionally used on light-

duty vehicles (Jiménez-Palacios, 1999) and its

generic definition, which includes the forces applied

to a moving body, is presented in Equation 1. The

coefficients of the equation are adjusted according to

the typology of vehicle monitored (Baptista et al.,

2013a). In this case, the coefficients used, adapted to

typical utility bicycles based on literature values

(Wilson, 2004), are presented in Table 3, obtaining

the bicycle specific power (BSP).

VSP







E





F



∙

F



∙



∙



∙



1





∙









∙



(1)

Table 3: Coefficient values for the variables included in

BSP.

Variables Values



0.01

g (m/s

) 9.81

0.008

1.2

A (m

) 0.5

cle

(kg) 18

ike

(kg) 70



a (kg/m

)

1.2

Using the respective coefficients, the BSP, in

W/kg) is defined by Equation 2:

   ∙



1.01∙ 9.81∙







0.078



0.0041∙



(2)

Similarly to the VSP methodology, the BSP is also

divided in modes that cover the full spectrum of the

bicycle operation, according to the following

formulation: group points with similar BSP values

(in W/kg); each BSP mode must include more than

1% of the total trip time, providing

representativeness for each mode; and the number of

modes is such that the total trip time is not

concentrated in a limited number of points.

Table 4 presents the modes (or bins) used in this

work and the respective range of power per mass.

Table 4: Binning method for BSP.

BSP mode Definition BSP mode Definition

<-4 BSP<-1 1 0≤BSP<1

-4 -4≤BSP<-3 2 1≤BSP<2

-3 -3≤BSP<-2 3 2≤BSP<3

-2 -2≤BSP<-1 4 3≤BSP<4

-1 -1≤BSP<0 >4 BSP>4

0 BSP=0

The percentage of time spent in each BSP mode

for the conventional and electric bicycles is

InfluenceofElectricBicycleUsageonBikerEffort-On-roadMonitoringApplicationinLisbon,Portugal

269

presented in Figure 2 and Figure 3. For negative

modes, the driving profile is very similar for both

bicycles. However, on positive BSP modes, the

electric bicycle present a higher share of time spent

in high BSP modes (higher power demands). This is

due to the electric assistance, which allows traveling

at high speeds on higher slopes and combinations of

higher speeds and acceleration, etc.

Figure 2: Time distribution (%) per BSP mode for

conventional bicycle.

Figure 3: Time distribution (%) per BSP mode for the two

electric bicycles.

Figure 4 presents the energy rate spent at each

BSP mode, on average, for the electric bicycles

studied, using the 1 Hz data from voltage and

current provided by the battery, measured under on-

road conditions. As expected, the energy rate

increases with BSP mode, showing the dominance

of electric assist on these modes.

Figure 4: Electricity consumption for the electric bicycles

as a function of BSP.

Although the data presented so far allows taking

conclusions about usage patterns of biker in EB and

CB, in both electric and conventional bicycles it is

necessary to determine the physiological impacts of

each technology and the respective human energy

expenditure to address the total energy impacts. The

Results section focus on the methodology developed

to assess this crucial part of the study.

3 RESULTS

Using the monitored heart rate data, the objective

was to correlate trip dynamic variables (traduced by

BSP) with HR variations. Hence, since BSP

aggregates trip information of speed, acceleration

and slope, its influence on heart rate was analyzed.

Additionally, since HR differ from person to person,

this analysis was done considering its derivate,



∆

∆

, and not its absolute value. Also, since HR is

an indirect unit of energy (as verified in the

Introduction), 

∆

∆

 traduces the variation of

Energy in a period of time, hence a measure of

Power.

Due to the existence of some noise in the HR

signal due to movement and vibration, the HR and

BSP second by second data was aggregated in a

minute by minute basis. In total, over 8 hour of data

was collected.

Figure 5 presents a clear relation between BSP

and 

∆

∆

. For mode 0, that corresponds to the biker

stopped, a reduction in 

∆

∆

 is observed, which

means that the biker reduced his HR in this

condition, compared with the previous riding

condition (thus 

∆

∆

0. For positive BSP

modes, which require more power from the biker,

positive variations in HR are observed due to the

increased human effort. As a result, for increasing

BSP modes, increasing positive variation in 

∆

∆



are observed. BSP mode >4 has few riding data

points, which justifies its divergence. For negative

BSP modes, which usually correspond to braking or

descent situation, reductions in HR/s are observed.

Figure 5: Influence of BSP in (∆HR/∆t) for the total data

collected.

PhyCS2014-InternationalConferenceonPhysiologicalComputingSystems

270

The collected data was also disaggregated

according to conventional and electric bicycle.

Figure

presents the 

∆

∆

 average results according to the

usage of conventional or electric bicycle. Over

4 hours of data are represented both for electric and

conventional bicycles (considering the average of

two electric and conventional bicycles that were

monitored). The variations in HR/s are lower for the

electric bicycle. That is mainly visible in positive

BSP modes, where higher power demand is

observed. For the electric bicycle, having the electric

motor assistance, helps reducing the human effort

and, consequently, variations in HR/s are lower. For

negative BSP modes, that difference is not so

visible, with both bicycles leading to reductions in

HR. It should be noticed that for the highest BSP

mode, the trend is not followed due to the lack of

points to fully characterize those conditions.

Figure 6: Influence of BSP in (∆HR/∆t) for conventional

(gray) and electric bicycles (black).

In order to obtain an average HR variation at the

end of the trip, the temporal distribution of BSP was

multiplied by the HR variations. Table 5 presents the

average

∑

∆HR



for each type of bicycle. The

electric bicycle leads to lower values compared to

the conventional one, with an average 57%

reduction.

Table 5:

∑

∆HR



for CB and EB.

Bicycle





∆



EB 0.59

CB 1.37

The next step was to analyze the human energy

expenditure (EE) associated to each trip. This

corresponds to the energy spent by the biker to drive

the bicycle. According to the literature review

(Ainslie et al., 2003); (Gastinger et al., 2012);

(Keytel et al., 2005) this corresponds to an accurate

approximation to account with the energy the body

burns during physical activity. The estimation of EE

can be performed using the equations presented in

Table 6.

Table 6: Quantification of EE from HR.

Source EE

(Keytel et al.,

2005)

EE gender



55.09690.6309HR

0.1988 weigh

0.2017age







1gender







20.4022 0.4472

HR 0.1263 weight

0.074age



EE in kJ/min;HR in BPM; gender=1 for man and 2

for woman. In this study, all biker were male, with an

average age of 28 and average weight of 70 kg.

(Gastinger et

al., 2012)

EE  0.103HR 4.795

EE in kcal/min;HR in BPM

(Ainslie et al.,

2003)

EE  0.0056 HR



0.6908HR 26.532

EE in kJ/min;HR in BPM

Average

equation

 













.

.

The 3 equations presented can be represented

simultaneously to obtain Figure 7 and an average

equation was obtained (Table 6) that was used for

the purpose of this study.

Figure 7: EE as a function of HR.

Using the 8 hours of physiologic data (divided in

4 hours for CB and the remaining for EB) and

recurring to the obtained average equation (Table 6),

the EE value was estimated to each second of the

trip. The next logical step was to group points with

similar BSP conditions, to obtain a representative

EE value associated to each BSP mode. This data is

presented in Figure 8, with the gray bars

representing CB and the black bars representing the

EB.

Figure 8: EE per BSP mode for CB and EB.

Figure 8 presents the results of multiplying the EE

profiles (in Wh/s) by each trip BSP time distribution,

for EB and CB. This way, the EE (in Wh) for each

BSP mode is obtained. Adding the EE at each BSP

InfluenceofElectricBicycleUsageonBikerEffort-On-roadMonitoringApplicationinLisbon,Portugal

271

mode and dividing by the trip distance, an estimate

of the human energy expenditure per kilometer

(Wh/km) is assigned for each trip and technology

used (conventional or electric).

In order to obtain the total energy consumption

(human and electric), an approach similar to the one

described previously was used for estimating electric

use, according to the consumption profile

distribution from Figure 4, instead of Figure 8.

Figure 9 presents an estimate of the total energy

per kilometer (human plus electric in electric bicycle

and human only in conventional bicycle). For the

conventional bicycle the total energy is around

69.8 Wh/km, while the electric bicycle presents

50.6 Wh/km of human energy (less 27.5% compared

with conventional bicycle) and 9.2 Wh/km of

electric consumption, resulting in a total of

59.8 Wh/km. Therefore, the total energy per

kilometer is 14.3% lower in the electric bicycle than

in the conventional.

Figure 9: Total energy expenditure for CB and EB.

With the data collected, it was not possible to

effectively estimate the efficiency of the electric

motor and the human body while cycling. However,

it is possible to obtain a set of acceptable values for

those efficiencies. Therefore, it was assumed that to

travel the distance of one kilometer it is necessary

the same energy, independently of using the electric

and conventional bicycle (Eq. 3).











EE





(3)











EE











Electricit

consumption







E





Where E

_

is the required energy for CB;

_

is the required energy for EB; ε



– is the

human body efficiency; ε



is electric motor

efficiency; EE



is the EE for CB; and EE



is the

EE for EB. Using to Eq 3 and assuming a range of

typical efficiency values for an electric motor,

between 60 to 95%, the range of human efficiency

can be estimated, as presented in Figure 10. As a

result, while cycling, the efficiency of human body,

theoretically, ranges from 30% to 45%.

Figure 10: Human efficiency versus electric motor

efficiency.

4 CONCLUSIONS

This research work addressed the use of

conventional and electric bicycles in real world

conditions, in order to estimate its impacts on

physiological signals, in more detail, in heart rate

and human energy expenditure. For this purpose a

methodology to express the power required to

overcome a drive cycle was adapted for bicycles,

resulting in the BSP methodology. The application

of this methodology used as basis data from the

monitoring of 6 bikers using both CB and EB, over

114 km in the city of Lisbon, Portugal, showing that

EB allow reaching higher BSP modes. However, the

developed methodology is not city or route specific

and can be applied elsewhere.

The impact on heart rate from shifting from

conventional bicycle usage to electric bicycle usage

was estimated. An average 57% reduction in HR

variations was found for the use of EB in typical

trips since, in high BSP modes that represent power

demanding situation, the electric motor comes in

action, avoiding human effort.

Moreover, a methodology was developed to

quantify the energy expenditure, based on heart rate

data measured under regular bicycle operation,

associated to the human effort that results from

using the bicycles. For the conventional bicycle the

total human energy expenditure reaches 70 Wh/km,

while the electric bicycle presents 51 Wh/km of

human energy (27.5% lower than the conventional

bicycle) and 9 Wh/km of electric consumption,

resulting in a total of 60 Wh/km. Consequently, the

total energy per kilometer is 14.3% lower in the

electric bicycle compared to the conventional.

In all, an innovative method of quantifying the

benefits for the biker of using electric bicycles was

developed, resulting in significant reduction in heart

rate variations, as well as, considerable energy

efficiency improvements. Using all the modal

information from Figures 4 and 8 regarding electric

PhyCS2014-InternationalConferenceonPhysiologicalComputingSystems

272

and human energy rates, combined with BSP modal

distribution for any route (as is presented in Figures

2 and 3), this methodology allows estimating the

total energy expenditure (human and electric),

electric autonomy, as well as HR variations,

according to the trip profile. As a result, this

methodology can be applied to evaluate the potential

use of EB in specific situation, namely bike-sharing

routes, providing significant support to bike-sharing

systems design and deployment.

ACKNOWLEDGEMENTS

The authors would like to acknowledge the sponsors

of the research: Prio.e and Eco-critério. Thanks are

also due to Fundação para a Ciência e Tecnologia

for the PhD and Post-Doctoral financial support

(SFRH / BPD / 79684 / 2011, SFRH / BPD / 62985 /

2009).

REFERENCES

Ainslie, P., Reilly, T., Westerterp, K., 2003. Estimating

Human Energy Expenditure, A Review of Techniques

with Particular Reference to Doubly Labelled Water.

Sports Med, 683-698.

Baptista, P., 2013. On-road Monitoring of Electric

Bicycles and its use in Bike-sharing Systems, in:

Silva, C. (Ed.), Grid Electrified Vehicles:

Performance, Design and Environmental Impacts.

Nova Science Publishers, Inc.

Baptista, P., Duarte, G., Mendes, M., Rolim, C., Farias, T.,

2013a. Scenarios for electric bicycle use: from on-road

monitoring to possible impacts of largeintroduction,

NECTAR Conference on Dynamics of Global and

Local Networks, São Miguel Island, Azores (Portugal).

Baptista, P., Pina, A., Duarte, G., Rolim, C., Pereira, G.,

Silva, C., Farias, T., 2013b. On road evaluation of

electric and conventional bicycles for urban mobility.

submitted to Transport Policy.

Barclays Cycle Hire, 2012. Barclays Cycle Hire -

Bikesharing in London.

Buehler, R., Pucher, J., 2012. Walking and Cycling in

Western Europe and the United States - Trends,

Policies, and Lessons. TR News May-June 2012:

Active Transportation: Implementing the Benefits.

Callabike, 2012. Callabike bike sharing system.

Cap'Vélo, 2012. Se déplacer à vélo en Grand Poitiers -

Bikesharing scheme in Saint-Etienne.

Cherry, C., Cervero, R., 2007. Use characteristics and

mode choice behavior of electric bike users in China.

Transport Policy 14, 247-257.

Cherry, C., Weinert, J., Xinmiao, Y., 2009. Comparative

environmental impacts of electric bikes in China.

Transportation Research Part D 14, 281–290.

Dill, J., Rose, G., 2012. E-bikes and Transportation

Policy: Insights from Early Adopters, 91th Annual

Meeting of the Transportation Research Board.

Eco-critério, 2013. http://www.ecocriterio.pt/.

EUROSTAT, 2013. Environment and Energy, EUROPA

Eurostat – Data Navigation Tree,

http://ec.europa.eu/eurostat.

Freemark, Y., 2010. Transit Mode Share Trends Looking

Steady; Rail Appears to Encourage Non-Automobile

Commutes, The Transport Politic.

Galp Energia, 2012. Ciclovias e percursos pedonais.

Gardner, G., 2010. Power to the Pedals. World Watch

Magazine 23.

Gastinger, S., Nicolas, G., Sorel, A., Sefati, H., Prioux, J.,

2012. Energy Expenditure Estimate by Heart-Rate

Monitor and a Portable Electromagnetic-Coil System.

International Journal of Sport Nutrition and Exercise

Metabolism, 117 -130.

Gastinger, S., Sorel, A., Nicolas, G., Gratas-Delamarche,

A., Prioux; J., 2010. A comparison between

ventilation and heart rate as indicator of oxygen uptake

during different intensities of exercise. Journal of

Sports Science and Medicine, 110-118.

Jiménez-Palacios, J., 1999. Understanding and

Quantifying Motor Vehicle Emissions with Vehicle

Specific Power and TILDAS Remote Sensing.

Massachusets Institute of Technology.

Keytel, L., Goedeck, J., Noakes, T., Hiiloskorpi, H.,

Laukkanen, R., Van Der Merwe, L., Lambert, E.,

2005. Prediction of energy expenditure from heart rate

monitoring during submaximal exercise. Journal of

Sports Sciences, 289 – 297.

Lathia, N., Ahmed, S., Capra, L., 2012. Measuring the

impact of opening the London shared bicycle scheme

to casual users. Transportation Research Part C:

Emerging Technologies 22, 88-102.

Lindsay, G., Macmillan, A., Woodward, A., 2011. Moving

urban trips from cars to bicycles: impact on health and

emissions. Australian and New Zealand Journal of

Public Health 35.

Martinez, L., Caetano, L., Eiró, T., Cruz, F., 2012. An

optimisation algorithm to establish the location of

stations of a mixed fleet biking system: an application

to the city of Lisbon, 15th Edition of the Euro Working

Group on Transportation, Paris.

Parkin, J., 2011. The importance of human effort in

planning networks, NECTAR Workshop Abstract.

Prio Energy, 2012. Prio Energy - Electric Mobility.

Pulkkinen, A., Saalasti, S., Rusko, H., 2005. Energy

expenditure can be accurately estimated from HR

without individual laboratory calibration, 52nd Annual

Meeting of the American College of Sports Medicine,

Nashville, Tennessee , USA.

The Bike-sharing Blog, 2011. Japanese Bike-share is

Electric.

Urban Audit, 2012. Urban Audit: Welcome to the Urban

Audit web site.

Vélib, 2012. Vélib - Bikesharing scheme in Paris.

VéliVert, 2012. VéliVert - Bikesharing scheme in Saint-

Etienne.

InfluenceofElectricBicycleUsageonBikerEffort-On-roadMonitoringApplicationinLisbon,Portugal

273

Wang, R., 2011. Autos, transit and bicycles: Comparing

the costs in large Chinese cities. Transport Policy 18,

139–146.

Wilson, D.G., 2004. Bicycling Science. The MIT Press.

Yu, Z., Völgyi, E., Wang, R., Ember, A., Wiklund, P.,

Alén, M., Tylavsky, F., Cheng, S., 2012. Comparison

of heart rate monitoring with indirect calorimetry for

energy expenditure evaluation. Journal of Sport and

Health Science, 178-183.

PhyCS2014-InternationalConferenceonPhysiologicalComputingSystems

274